When submitting a Service Request (SR), please include an archive file with the relevant log files and debugging information as listed in this section. This information can be used by Oracle Support to analyze and diagnose system issues. The support data files can be uploaded for further analysis by Oracle Support.
Collecting support files involves logging in to the command line on components in your Oracle PCA rack and copying files to a storage location external to the appliance environment, in the data center network. This can only be achieved from a system with access to both the internal appliance management network and the data center network. You can set up a physical or virtual system with those connections, or use the master management node.
The most convenient way to collect the necessary files, is to mount the target storage location on the system using nfs, and copy the files using scp with the appropriate login credentials and file path. The command syntax should be similar to this example:
# mkdir /mnt/mynfsshare# mount -t nfsstorage-host-ip:/path-to-share/mnt/mynfsshare # scp root@component-ip:/path-to-file/mnt/mynfsshare/pca-support-data/
Collecting Oracle Private Cloud Appliance Support Data
For more accurate diagnosis of physical server issues,
Oracle Support Services require a system memory
dump. For this purpose, kdump must be
installed and configured on the component under investigation.
The entire procedure is described in the support note with
Doc
ID 1520837.1. By default, kdump
is installed on all Oracle PCA compute nodes and configured
to write the system memory dump to the ZFS storage appliance
at this location:
192.168.4.1:/export/nfs_repository1/.
For diagnostic data collection, Oracle Support Services recommend that the OSWatcher tool be run for an extended period of time. For details about the use of OSWatcher, please consult the support note with Doc ID 580513.1. OSWatcher is installed by default on all Oracle PCA compute nodes.
For diagnostic purposes, Oracle Support Services use a tool called pca-diag that automatically collects vital troubleshooting information from your Oracle PCA environment. This tool is part of the Oracle PCA controller software installed on both management nodes and on all compute nodes. Its capabilities are described in the section “Oracle Private Cloud Appliance Diagnostics Tool” in the Oracle Private Cloud Appliance Administrator's Guide.
To collect support data from your system, proceed as follows:
Log in to the master management node as root.
Run pca-diag with the appropriate command line arguments.
NoteOracle Support teams may request that the tool be run in a specific manner as part of an effort to diagnose and resolve reported hardware or software issues.
For the most complete set of diagnostic data, run the command with both arguments: pca-diag ilom vmpinfo.
pca-diag ilom
Use this command to detect and diagnose potential component hardware and software problems.
[root@ovcamn05r1 ~]# pca-diag ilom Oracle Private Cloud Appliance diagnostics tool Gathering Linux information... Gathering system messages... Gathering PCA related files... Gathering OS version information... Gathering host specific information... Gathering PCI information... Gathering SCSI and partition data... Gathering OS process data... Gathering network setup information... Gathering installed packages data... Gathering disk information... Gathering ILOM Service Processor data... this may take a while Generating diagnostics tarball and removing temp directory ============================================================================== Diagnostics completed. The collected data is available in: /tmp/pcadiag_ovcamn05r1_
<ID>_<date>_<time>.tar.bz2 ==============================================================================pca-diag vmpinfo
Use this command to detect and diagnose potential problems in the Oracle VM environment.
NoteTo collect diagnostic information for a subset of the Oracle VM Servers in the environment, run the command with an additional
serversparameter, as shown in the example below.[root@ovcamn05r1 ~]# pca-diag vmpinfo servers='ovcacn07r1,ovcacn08r1' Oracle Private Cloud Appliance diagnostics tool Gathering Linux information... Gathering system messages... Gathering PCA related files... Gathering OS version information... Gathering host specific information... Gathering PCI information... Gathering SCSI and partition data... Gathering OS process data... Gathering network setup information... Gathering installed packages data... Gathering disk information... Gathering FRU data and console history. Use ilom option for complete ILOM data.
When the
vmpinfo3script is called as a sub-process frompca-diag, the console output continues as follows:Running vmpinfo tool... Gathering files from servers: ovcacn07r1,ovcacn08r1 This process may take some time. The following server(s) will get info collected: [ovcacn07r1,ovcacn08r1] Gathering OVM Model Dump files Gathering sosreport from ovcacn07r1 Gathering sosreport from ovcacn08r1 Gathering OVM Manager Logs Gathering manager sosreport.
When all files have been collected, the data is compressed into two tarballs. One is from the
pca-diagtool, whilevmpinfo3writes a separate tarball with its own specific data.Compressing VMPinfo3
<date>-<time>. ======================================================================================= Please send /tmp/vmpinfo3-<version>-<date>-<time>.tar.gz to Oracle OVM support ======================================================================================= Generating diagnostics tarball and removing temp directory ============================================================================== Diagnostics completed. The collected data is available in: /tmp/pcadiag_ovcamn05r1_<ID>_<date>_<time>.tar.bz2 ==============================================================================
If necessary, run pca-diag, with or without the
ilomargument, on some or all compute nodes as well.To allow better analysis of physical server issues, for example hanging, crashing or rebooting, also include the system memory dump file (
vmcore).The location of the file is:
<kdump-partition-mount-point>/var/crash/127.0.0.1-<date>-<time>/vmcorekdumpconfiguration. By default,kdumpwrites to192.168.4.1:/export/nfs_repository1/. For details, please consult the support note with Doc ID 1520837.1.When required, collect the OSWatcher logs from the compute nodes. The default location is
/opt/osw.For details, please consult the support note with Doc ID 580513.1.
Copy all diagnostic files to a location external to the appliance environment.
Uploading Support Data Files
For support data up to 2 GB, upload the file as part of the Service Request (SR) process in My Oracle Support (MOS).
If you are still in the process of logging the SR, upload the support data in the Upload Files/Attachments step of the SR.
If you have already logged the SR and need to upload files afterwards, proceed as follows:
Log into MOS and open the Dashboard or Service Request tab.
In the Service Request region, click the SR you want to update.
In the Update section, select Add Attachment.
In the pop-up window, select the file for upload, include any notes, and click Attach File.
If uploading the support data with the SR is not an option, or
for support data files over 2 GB in size, use the FTPS file
upload service from Oracle support at
transport.oracle.com. Oracle
Support might request that you upload using a different
mechanism.
Using an FTPS client, for example FileZilla or WinSCP, access the My Oracle Support File Upload Service
transport.oracle.comin passive mode.Log in with your Oracle Single Sign-On user name and password.
Select the support data file to upload.
Select a destination for the file.
Use the directory path provided by Oracle Support.
Typically, the directory path is constructed as follows: "
/upload/issue/".<sr_number>/The use of a SR number ensures that the file is correctly associated with the service request. Write down the full path to the file and the SR number for future reference in communications with Oracle Support.
Upload the file.
When the upload is complete, a confirmation message is displayed.
If you prefer to use a command-line client, for example cURL,
you typically enter a single command to connect, authenticate
and complete the upload. A cURL command will look similar to
this example: curl -T
.
For security reasons, it is recommended that you leave the
password out of the command and be prompted to enter the
password.
<path_to_file> -u
"<user>"
ftps://transport.oracle.com/upload/issue/<sr_number>/
For detailed information about uploading files to Oracle Support, refer to the support note with Doc ID 1547088.2.